The invention relates generally to a method and an engine for evaluating a service based on a request for a service, in particular an information technology service. The information technology service may in particular be dedicated to a Cloud Computing environment. The invention relates further to a service evaluation engine, a computer system, data processing program, and a computer program product.
In the current days of Cloud Computing, i.e., consuming information technology services (IT-service) as a service, either from a Public Cloud Computing environment or from a Private Cloud Computing environment, it is required to manage virtual machines, configurations and infrastructures in a most efficient and dynamic way. In this context, a Private Cloud Computing environment may be viewed as applying the rules of a Public Cloud Computing environment and apply it to enterprise data centers. Cloud services may be defined, as shared services under a virtualized management, being accessible over the Internet by people or other services via Internet standards.
A creation of dynamic IT-service templates is a central task in modern data centers to allow data center customers requesting IT services based on these service templates. Templates may be seen as generic predefined cook-book-like definitions of classes of resources that may not be specified on the most detailed level possible, but on a generic level adhering to classes of resources.
Thereby, a multitude of different service components may lead to a vast set of possible configurations, each of them contributing to various quality-of-service aspects, e.g., performance, availability, security, price, etc.
An automatic management or only a comprehensible management of required quality-of-service aspects may hardly be automated and may require often expert knowledge of an experienced data center employee. This may make an automatic management of IT-services, particularly in Cloud Computing environments, hardly achievable.
Several approaches have been followed in order to improve drawbacks of automatic or semi-automatic resource configuration in IT environments. One approach has been disclosed in CN11383854. A method for providing WebServices based on a user preference is disclosed. The method comprises the following steps: establishing a WebService composition template library which at least comprises a WebService composition template profile as well as a list and a process for abstract services, which are all described by an ontology concept in an ontology library; ensuring the preference degree of a user to the ontology concept; and searching a WebService composition template which is matched with a WebService request according to the WebService request of the user.
Today's rule-based validation systems may evaluate each possible configuration of dynamic IT-service templates to either “valid” or “invalid” instead of being based on preference degrees. There may only be a volume but no value business support, meaning that masses of configurations may be validated in terms of technical validation and functionality but not in terms of business requirements.
Additionally, these systems may not include abstraction levels like, e.g., a class named “Linux” abstracting from the various Linux distributions. That is, for each possible configuration instance there has to be an individual rule. With a growing number of different resources this may be an unmanageable approach.
Neither do the existing systems have the possibility to infer quality-of-service aspects, based on known patterns of atomic entities, nor do they allow higher level class relationships.
Typically, the set of rules may closely be intertwined with program code of generation tools that may be used for generating the rules and, therefore, hard to maintain and evolve.
However, because of the above-mentioned drawbacks and available technology, there may be a need, for a flexible and dynamic evaluation of services, particularly IT-services relating to quality-of-service (QoS) aspects of a request for an IT-service. It may also be advantageous having a mechanism for an automatic selection of resources relating to QoS requirements if several options to fulfill a request may be available.